KR102121089B1 - Systems, devices and methods for the quality assessment of oled stack films - Google Patents

Abstract

The present invention provides a technique for evaluating the quality of a deposited film layer of an organic light emitting diode ("OLED") device. Images are captured and filtered to identify the deposited layer to be analyzed. The image data representing this layer can optionally be converted into brightness (grayscale) data. Then, a gradient function is applied to highlight discontinuities in the deposited layer. Then, the discontinuity is compared to one or more thresholds to check the quality of the deposited layer, and an optional remedy is applied. The disclosed technique can be applied in place, so that potential defects, such as interlayer delamination, can be quickly identified before subsequent manufacturing steps are applied. In an optional embodiment, the remedy may vary depending on whether a defect is present.

Description

SYSTEM, DEVICES AND METHODS FOR THE QUALITY ASSESSMENT OF OLED STACK FILMS}

The present invention claims priority to US Provisional Patent Application No. 61/766,064 filed February 18, 2013 for "Systems and Methods for the Detection of Defective EML Films" on behalf of the first inventor Cocoa Christopher And the priority patent application is incorporated herein by reference.

The teachings of the present invention relate to systems, devices, and methods useful during the manufacture of organic light emitting diode ("OLED") devices for the quality assessment of various films formed within pixel well structures.

Most attention to the potential of OLED device technology was driven by the emergence of a very thin, energy-efficient flat panel with a high-saturated color. In addition, a wide variety of substrate materials, including flexible polymer materials, can be used in the fabrication of OLED devices. OLED devices can be manufactured by printing a variety of organic and other thin films on a substrate using an industrial printing system. From substrates sized for use as cell phone displays to substrates sized for use as very large television ("TV") screens, substrates of almost any size can be used in this process. To provide two non-limiting examples, ink jet printing of thin films can be used on Gen 7.5 substrates, which have dimensions of about 195 cm x 225 cm, which substrates are 8 42" or 6 47 per substrate. "It is cut into flat panels, for Gen 8.5 substrates, with dimensions of 220 x 250 cm, these substrates are cut into six 55" or eight 46" flat panels per substrate.

OLED devices typically have a plurality of pixels to construct a display. In color displays, each pixel typically has three separate color generating elements. Eventually, each of these devices typically uses a “well” to accommodate one or more thin film layers during the ink jet printing process. Thus, each pixel of the OLED device is typically associated with three wells corresponding to each pixel color. The set of layers for each color scheme (ie, associated with each well) forms the “OLED stack” dmf. Each OLED stack can include 6-7 film layers. During manufacture, it is desirable to deposit each of these layers uniformly.

Visually, a high-definition flat panel display has a pixel density of about 300 ppi to about 450 ppi, and may include more than 2 million pixels. Obviously, considering the sheer number of functional pixels that must be formed on the substrate during the manufacture of various OLED devices, a high degree of manufacturing accuracy is required. In the process of forming various layers, various discontinuities between or within the film layers may occur, which may not perform as specified, or may result in identification as a defect.

Accordingly, there is a need in the art for systems, devices and methods that can be used to properly and systematically evaluate the quality of thin films formed on substrates during the manufacture of OLED devices.

The illustrative embodiments of the present invention will be described below with reference to the accompanying drawings.
1A is a schematic diagram of an exemplary pixel arrangement in a display panel, in accordance with the present invention. 1B is a schematic diagram of an embodiment of an OLED stack, in accordance with the present invention.
2A is a cross-sectional view illustrating an illustrative pixel well, in accordance with the present invention. 2B is a top view showing a structure associated with a single pixel, according to the present invention.
4 shows a block diagram of a data collection device of the panel inspection system of FIG. 3, according to an illustrative embodiment.
5 shows a printing system with the data collection assembly of FIG. 4, according to an illustrative embodiment.
6 is a schematic cross-sectional view of a gas enclosure system capable of accommodating a printing system, such as various embodiments of the printing system of FIG. 5.
7A-7D illustrate various flowcharts illustrating exemplary tasks performed by image processing applications of a data collection device, in accordance with various embodiments of the systems and methods of the present invention.
8A-8F show schematic diagrams of one or more pixel wells, used to represent the work of an image processing application, in accordance with various embodiments of the present invention.
9 shows a histogram of gradient intensity corresponding to a view of the layer deposited in the pixel well (viewable from the left side of FIG. 9).
FIG. 10 shows a histogram of the gradient intensity corresponding to a view of the layer enhanced in the pixel well (viewable from the left side of FIG. 10 ).
11 is a flow chart illustrating exemplary operations performed by an image processing application of a data collection device, in accordance with various embodiments of the present invention.
12 is a flow chart illustrating an exemplary job performed by an image processing application of a data collection device, in accordance with various embodiments of the present invention.

The present invention provides a system, apparatus and method for evaluating the quality of a thin film layer deposited during OLED device fabrication. One or more layers of the OLED stack can be continuously printed on a target area of the substrate, each target area being a pixel well that will optionally be associated with a particular color scheme of pixels of light to be produced by the finished OLED device. The inkjet printing process can optionally be used for this printing process. The deposited layer may be formed of an organic or inorganic material, but typically, the OLED stack includes at least one organic layer (ie, a layer of shedding material or “EML”) formed using this process. After printing a specific ink into the targeted area, one or more post-printing processing steps can be performed to complete each layer, such as by converting the deposited fluid to a permanent structure. To evaluate the quality of each layer formed within each target area, images of all target areas are captured during or after layer deposition and/or formation, such as using a high-speed, high-resolution camera. Such imaging can optionally be performed prior to deposition of the next layer of the OLED stack, to evaluate the previous'wet' layer or the finished layer (ie at any stage of the layer formation process).

Non-uniformity in the deposited layer can be detected through evaluation of image data taken as a result of this image capture. Each captured image is typically a high quality close-up view of one or more pixel wells or pixels of one or more OLED device substrates. Non-uniformity can be expressed as a discontinuity between layers of film in a pixel well, which indicates interlayer peeling between films, gaps, pinholes, or other types of problems. Each captured image is a non-limiting example, one or more pixel wells, an area surrounding one or more pixel wells, a bank forming a confinement of each pixel well, and each for forming part of an OLED stack. It may include a film deposited in a pixel well. The captured image separates the generated filtered data, separates the image data directly corresponding to the deposited layer of the film of interest, and removes excess data (eg, image data for a given pixel well or an outer region of wells). To be filtered. This filtered data is typically the very image data of the film layer in question. Gradient functions can be applied to these filtered data to form processed data. Processed data is typically a gradient valued image that emphasizes discontinuities from uniform image data. For various embodiments of the systems and methods of the present invention, such processed data is used to evaluate the quality of the OLED stack film in one or more wells, depending on, for example, the size of the discontinuities, the number of discontinuities, or one or more other criteria. Can be used. Then, a result or output representing the quality of the pixel wells that were evaluated using the processed image data can be generated. In various embodiments of the present invention, this output may indicate whether or not a layer deposited within a particular pixel well has a fill problem or an interlayer peeling problem. In another embodiment, the described process is unacceptable, and can be applied repeatedly for each layer in the OLED stack in the pixel well, with results indicating errors used to influence subsequent processing. Finally, if any defects are identified, a remedy can be taken as an option. Obviously, the manufacturing process is not perfect, and timely detection of defects and the use of these remedies are important to maximize quality and speed of production and minimize cost.

In one embodiment, the filtered data (i.e., image data representing the layer deposited under analysis) is such as erection, grayscale value, hue, color intensity, or other image features (before application of the gradient function). It can be converted to a specific format to help emphasize specific image features. In one embodiment, for example, filtered data representing a color image is converted to an 8-bit grayscale intensity value, one for each pixel or "PEL" of the filtered data ("PEL" is typically captured by the camera). Whereas “pixels” will be used to refer to pixels of high resolution data, “pixels” will typically be used to refer to the photographic elements of the finished OLED panel and the associated photogenerating configurations and/or areas occupied by these configurations). Then, this emphasized data (i.e., monochrome image data in the case of grayscale conversion) is generated with processed data by receiving a gradient function. The gradient function, as mentioned, emphasizes non-uniformity on a local basis within the pixel wells under scrutiny.

One embodiment of the present invention provides a computer-readable medium comprising computer-readable instructions stored on a computer-readable medium. When executed, these computer-readable instructions cause the processor to process captured image data representing a target area for analyzing the quality of the deposited film. In other words, the target region may include at least one pixel well represented by the captured image. Computer-readable media are non-transitory media, meaning that they are physical structures suitable for storing data in electrical, magnetic, optical, or other forms. Examples of such media include, for example, flash drives, floppy disks, tapes, server storage or mass storage, hard drives, dynamic random access memory (DRAM), compact disks (CD), or other local or remote storage. The computer-readable medium can have instructions stored therein, when executed by a processor, cause the system to perform various embodiments of methods for identifying discontinuities in films formed on OLED device substrates. The computer-readable medium may alternatively be implemented as part of a larger machine or system, or may exist on a separate basis (i.e., a flash drive or standalone storage to transfer files to another computer or printer later). The larger machine or system may optionally include a printing device including a camera and processor and optionally a printhead and/or camera transport mechanism. That is, a camera that takes a picture of an inactive OLED device substrate during fabrication using illumination outside the substrate can be mounted on an assembly that includes a light source to illuminate a target area of the OLED device substrate during image capture.

In a more detailed embodiment, a gradient function is applied by a mathematical process to image data representing a specific target area, as a convolution between elements from two mattresses. Such image data may be the above-mentioned filtered data, and typically includes image data derived from PELS of the captured image. Such image data mentioned can highlight certain image features such as color intensity, brightness, and the like. Convolution may optionally be performed by software or firmware, ie, one or more machines operating under the control of software and/or firmware. The result of convolution can be expressed as a matrix of gradient values corresponding to the film layers under close scrutiny. Then, a measurement is applied (ie one or more criteria or thresholds) to evaluate the material of the deposited film. If the result satisfies the applied measurement(s), the deposited layer is determined to have a satisfactory quality, or it is determined that a potential problem exists. As a non-limiting example of an action that can be taken in response to an acknowledgment of a problem, the display panel can be rejected without further processing (potentially saving manufacturing time and cost), or additional processing may be applied in an effort to cure the problem. Can be. Note that a wide variety of different criteria can be applied to assess quality. For example, in one particular detailed embodiment discussed below, the absolute value of the gradient value for a layer in a pixel well is summed (or the square of the gradient value is summed), and if the sum result is less than the threshold, deposited The film is determined to be acceptable. In another embodiment, a second test may instead be applied to distinguish between a large and noticeable discontinuity or gradient (e.g., a significant problem) and a small discontinuity of complement (e.g., not a problem). Statistical tests can also be applied (e.g., the comparison of the standard deviation of values from the result matrix with the second threshold) to identify any specific features that may indicate a problem. For example, in one application discussed below, the histogram is computed for the gradient values associated with each pixel well containing the layer under scrutiny, and the high degree of occurrence of strong gradients results in interlayer delamination within a given pixel well. It can indicate "racetrack (racetrack)" associated with. In addition, these tests can be used in combination with one another if desired, such as if the layers in a given well are considered acceptable quality only if both tests are satisfied. Certainly, many alternative tests and combinations of tests also exist and can be selected for various embodiments.

In another embodiment, a second gradient function can be applied as an option. For example, the filtering process discussed above uses a gradient function to identify the contours of the deposited layers from the captured image, and these contours mask out portions of the captured image that do not directly represent the deposited layer under consideration. It is used to (mask out). In a more specific embodiment, each copy of the captured image is filtered using a gradient function, for example, to find a "bank" that structurally outlines each pixel well. The mask function is then defined based on this analysis, passing (ie, not masking) the image data in the bank (eg, which may correspond to the deposited layer of interest), and confinement of the problem pixel wells. The image data corresponding to the external substrate structure (ie, placed outside the bank) is blocked (ie, masked out). Then, this mask is applied to the original image data in order to obtain filtered data (i.e., representing layers directly under scrutiny).

Although apparent from above, the various operations, methods, and devices and systems that perform these operations discussed above enable timely and systematic evaluation of the quality of the films of OLED devices. Therefore, these various methods, devices and systems lead to more reliable OLED devices because potential defects can be detected more accurately. In addition, they lead to lower manufacturing costs, because these defects do not incur the cost and time associated with further processing of potentially defective OLED devices, and also the cost and time associated with the disposal of available materials. This is because it can be detected (and potentially cured).

Additional details and options will be apparent to those skilled in the art from the following discussion.

1A is a schematic top view of an OLED device substrate 910, including an enlarged view 20 showing a network for six pixels formed on the surface of the substrate 10. It can be seen that a single pixel is designated by the number 30 and consists of separate elements 32, 34 and 36 that generate red light, green light and blue light. Additional circuitry (such as designated by number 38) can be formed on the OLED device display substrate to assist control over the production of light by each pixel well.

As previously mentioned, during the manufacture of an OLED flat panel display, the OLED pixel is formed to include at least one OLED film stack capable of emitting light when voltage is applied. 1B is an OLED stack comprising an electron transport layer (ETL) coupled with an electron injection layer (EIL), and a hole injection layer (HIL), a hole transport layer (HTL), a divergence layer (EML), between the anode and the cathode. An example of a film structure is shown. When a voltage is applied across the anode and cathode, as shown in Fig. 1B, light of a specific wavelength is emitted from the EML layer. In various embodiments of the system, apparatus, and method of the present invention, the HIL, HTL, EML and ETL/EIL layers shown in FIG. 1B can be printed using inkjet printing. Each HIL, HTL and ETL.EIL OLED stack layer has an ink formulation comprising the materials that form the function of these OLED stack film layers. As will be discussed in more detail below, a pixel may include three color generating elements, each element emitting light of different wavelengths, such as, but not limited to, red, green, and blue light. It has an EML layer. For various embodiments of the OLED pixel cell of the present invention, each EML layer has an ink composition comprising an OLED material capable of emitting within a targeted electromagnetic wavelength range. Note that any number or combination of color configurations and associated pixel wells may have a monochrome display (eg, with a single pixel well for each pixel).

2A shows a cross-sectional view of a pixel well 50, in which ink for forming various OLED stack layers is printed to form a single color generating element. Although a cross section depicts a pair of double bank structures (designated by numbers 52 and 54) that serve to confine the deposited fluid, in fact, these structures are typically a single closed part of a shape (eg, each in FIG. 2B) Numbers 62, 64 and 66, as indicated by the contours of the red, green, or blue light-producing regions). In the case of the pixel well 50, the substrate 51 is typically a transparent structure, and consequently, the region 58 represents an area of active pixel width to be associated with the generated light (unobstructed), and the area ( 56) represents the well width associated with the deposited layer. The relative heights and distances of the various structures are confirmed by the legend seen in the upper right of FIG. 2A. As previously mentioned, in some embodiments, a pixel bank, more specifically a confinement bank 54, is used to filter and capture the captured image data used to analyze the deposited layer, so that only these pixel well widths ( The filtered image data corresponding to 56) is processed to evaluate the layer discontinuity. This is not required for all examples.

2B shows a top view of a pixel cell 60 with dimensions about 95 μ wide and about 85 μ long. In the pixel cell 60, there are three different wells 62, 64 and 66 each used to accommodate the layers forming the OLED stack, one stack 62 for a red light generating element, a green light generating element. One stack 64 for, and one stack 66 for blue light generating elements. The dimensions identified in Figure 2B confirm the pixel cell pitch and pixel well dimensions for each color element for a flat panel display (with a resolution of 326 pixels per inch or "ppi").

As one of ordinary skill in the art can readily recognize, the pixel size, shape, and aspect ratio can vary depending on a plurality of factors. For example, a pixel density equivalent to 100 ppi may be sufficient for a panel used for a computer display. In contrast, pixel densities ranging from 300 ppi to about 450 ppi can be used for ultra high resolution displays. Note that a wide variety of inks can be used to form different layers for use with light of various wavelengths, so that each well 62, 64 and 66 does not use the same inks or inks. In addition, various inks can be formulated to form other layers, as shown in FIG. 1B.

The present invention enables automated systems and methods for timely inspection and objective determination of the quality of a layer deposited over multiple pixel wells of an OLED device substrate. For example, it will be recalled that for many designs (eg, Gen 5.5 to Gen 8.5), there are typically millions of pixels per panel, reflecting the previously discussed discussion of various substrate sizes. This number can increase dramatically in the future, as additional display technologies are successful. As previously discussed, for proper function in OLED devices, film uniformity is desirable.

3 is a schematic diagram of an OLED device quality evaluation system 300 according to an illustrative embodiment. The quality evaluation system 300 may include a data storage system 305, as well as a data collection system 304, an image processing system 306, and a network 308. The data collection system 304 includes a camera, such as a high-speed CCD camera. The camera captures an image representing a high-magnification, high-resolution view of the OLED device substrate. In addition, the data collection system 304 can optionally be implemented with an OLED device fabrication mechanism, including one or more printers and/or other devices. In one embodiment, the camera is transported by a transport mechanism used to position the camera, as well as an ink jet printhead used to print various film layers. The captured image and/or other data can be stored in storage system 305. The image processing system 306 processes the captured image and determines if the pixel is defective or not, or does not perform as designed for utilization. Typically, the image processing system is a computer readable medium (ie, a portable device, a desktop cup or laptop or part of a laptop or other processing device shown) that allows one or more processors of the image processing system 306 to perform such an analysis. It includes. The network 308 connecting these various devices may include one or more networks of the same or different types. The network 308 includes, but is not limited to, a cellular network, a Bluetooth network, a P2P network, a local area network, a wide area network such as the Internet, an internal system bus, or another scheme for connecting various illustrated devices. Types of wired and wireless public or private networks may be included. Network 308 may also include subnetworks, and may consist of a number of additional or fewer devices. The configuration of the OLED device quality inspection system 300 can be included in a single computing device, can be located in a single room or adjacent rooms, a single facility, and can also be remote from each other. In various embodiments, the present invention may be implemented in any single configuration shown in FIG. 3, or as a set of such configurations, or as related methods, devices or systems.

4 shows a block diagram of a data collection device 400, according to an illustrative embodiment. The data collection device 400 can include a collection assembly 450, a camera assembly 460, a pointing system 470 and at least one processor 406 to provide one non-limiting example. . In order to assist input and output of data to at least one processor 406, the data collection device 400 may include an input interface 402, an output interface 404, a communication interface 408, a keyboard 410, a mouse ( 412), display 414, printer 416, computer-readable medium 420, and data collection application 430. Less, different, and additional configurations can be incorporated into the data collection device 400.

For the collection assembler 450 of the data collection device 400 of FIG. 4, the camera assembly 460 is optionally configured to capture image data using a stove, as understood by those skilled in the art. Can be. The captured image data is received into data collection device 400 via input interface 402. In addition, the data collection assembly 450 can include a pointing system 470, which mounts the camera assembly 460 and is used to transport the camera assembly to the OLED device substrate. In this regard, pointing system 470 provides for movement of camera assembly 460 over the entire surface of the flat panel display. Thus, the data collection assembly 450 of the collection device 400 can fully capture an image representing the entire set of pixel wells of any flat panel display device. As previously mentioned, in one embodiment, camera assembly 460 includes a printer (eg, printer 416) that includes a processor 406 and/or collection assembly 450 used to control print head movement. Note that it may be part of )). In such a system, camera assembly 460 can be integrated into the print head to capture an image of the just-deposited “wet ink” for purposes of quality analysis. Alternatively, the camera assembly 460 can optionally be implemented as part of a multi-tool printer, where the pointing system 470 is equipped with one or more printheads for use in printing OLED device substrates at some time. However, at other times, it performs a dual-role, which mounts the camera assembly 460. In one embodiment, a manufacturing apparatus for an OLED device transfers a plurality of substrates, so that after the layer is deposited on the first substrate in the printer, the substrate is advanced to another location in the manufacturing device, and imaged for quality analysis At the same time, a second substrate enters the printer for the purpose of depositing a similar layer.

The input interface 402 provides a port for receiving data into the data collection device 400, as understood by those skilled in the art. The input interface 402 is not limited, but interfaces with various input technologies including a keyboard 410, a mouse 412, a display 414, a camera assembly 460, a track ball, a keypad, one or more buttons, and the like. , Allows information to be received into the data collection device 400, or allows selection from a user interface displayed on the display 414. The keyboard 410 can be any of a variety of keyboards, as understood by those skilled in the art. The display 414 can be a thin film transistor display, a light emitting diode display, a liquid crystal display, or any of a variety of different displays as understood by those skilled in the art. The mouse 412 can be any of a variety of mouse-type devices as understood by those skilled in the art. The same interface can support both the input interface 402 and the output interface 404. For example, the display 414 can include a touch screen that supports user input and presents output to the user. The data collection device 400 may have one or more input interfaces using the same or different input interface technologies. The keyboard 410, mouse 412, display 414, camera assembly 460, and the like can be further accessed by the data collection device 400 through the communication interface 408.

The output interface 404 may provide an interface for outputting information from the data collection device 400. For example, the output interface 404 is not limited, but may interface with a variety of output technologies, including other devices that are not part of the display, printer, camera, pointing system, and data collection platform 400. In at least one embodiment, output interface 404 includes a local area connection or a wireless connection to communicate with other network-connected devices.

The processor 406 executes instructions as understood by those skilled in the art. Instructions may be performed by special purpose computers, logic circuits, or hardware circuits. The term "execute" refers to a process that performs the action required by each instruction. Instructions can be expressed in the form of one or more programming languages, scripting languages, assembly languages, and the like. The processor 406 receives instructions from a non-transitory computer-readable medium, such as local, network-connected or remote memory, stores copies of these instructions, such as in local RAM, and executes on these locally-stored instructions. By acting on, it executes these commands. In addition, the data collection device includes a plurality of processors, such as a multi-core device, a set of processors of an integrated device, or a network-connected device that collectively performs operations and exchanges data related to these operations. As a set of, it can perform various functions that are executed.

5 shows a front perspective view of an inkjet printing system 500 for printing various OLED stack layers on an OLED substrate, according to an illustrative embodiment. An OLED inkjet printing system, such as the inkjet printing system 500 of FIG. 5, can be comprised of several devices and devices, which are located at a specific location on the substrate, such as the substrate 522 positioned on the substrate support device 520. Reliable placement of ink drops. These devices and devices are not limited, but may include printhead assemblies, ink delivery systems, motion systems, substrate support devices and substrate loading and unloading systems. The movement system may be a gantry system or a split axis XYZ system as shown for the inkjet printing system 500 of FIG. 5 as a non-limiting example. The printhead assembly can move over a static substrate (Gentry style), or both the printhead and substrate can move (for split-axis configuration). In another embodiment, the printhead assembly may be substantially static, for example, the substrate may have one or more Z-axis movements provided by a substrate support device, or by a Z-axis movement system associated with the printhead assembly. It can be transported in X and Y dimensions for the printhead. As one or more printheads move relative to the substrate, droplets of ink are ejected at the correct time to deposit within a desired location on the substrate. The substrate can be inserted into or removed from the printer using a substrate loading and unloading system. Depending on the configuration of the inkjet printing system, this can be achieved using a mechanical conveyor, a substrate lift table with a conveyor assembly or a substrate transfer robot with an end effector. Note that in one embodiment, printing is performed in a sealed chamber against ambient air, excluding unwanted particulates. In addition, the data collection system is optional and can be operated in such a separate environment. In applications that may be considered, the sealed chamber, referred to below, can be used to perform printing and/or imaging in the presence of a specially selected atmosphere, such as an inert gas.

The inkjet printing system 500 of FIG. 5 can include a data collection assembly as previously described. Various embodiments of the inkjet printing system 500 may be supported by the printing system base 510. The first and second lifters 512 and 514 above this base support the bridge 516 used to support the printing system and the imaging system (eg, camera). The cable tray assembly 542 provides routing for the tubing assemblies required for the operation of the enclosure and various cables, wires and inkjet printing systems 500, and to include any particulates produced by these cables, wires and tubing assemblies help. The cable tray assembly 542 is generally part of the cable tray assembly exhaust system 540, which provides exhaust of these particulates through the tray exhaust plenum 544. In this regard, the cable tray exhaust assembly 540 can help enable a low-particulate environment during the printing and/or imaging process.

The bridge 516 helps support the first X, Z-axis carriage assembly 530 and the second X, Z-axis carriage assembly 530A. Each carriage assembly can move on the bridge 516 in the X-axis direction relative to the substrate support device 520. The inkjet printing system 500 can essentially use a low-particulate X-axis motion system, with the first and second X,Z carriage assemblies 530 and 530A moving on air bearing linear slider assemblies, respectively. Those of ordinary skill in the art can understand, and the air bearing linear slider assembly encloses the entirety of the bridge 516, allowing frictionless movement of the first and second X,Z carriage assemblies 530 and 530A. This air bearing linear slider assembly helps to provide three-point mounting, which preserves the accuracy of movement for each X,Z carriage assembly and helps to resist skew. The first X, Z-axis carriage assembly 530 has a first Z-axis motion plate 532, while the second X, Z-axis carriage assembly 530 has a second Z-axis motion plate 532A. And each Z-axis movement plate helps to control the Z-axis movement of the device mounted there. For the inkjet printing system 500, the printhead assembly 550 appears to be mounted on the Z-axis motion plate 532, while the camera assembly 560 is shown to be mounted on the Z-axis motion plate 532A. . In various embodiments of an inkjet printing system having two X,Z-axis carriage assemblies, it is easy to recognize that the printing assembly and camera assembly can be interchangeably mounted on each X,Z-axis carriage assembly. have. Other embodiments may use a single X,Z-axis carriage assembly with a printhead assembly mounted on a single X,Z-axis carriage assembly and a camera assembly mounted adjacent to the printhead assembly.

In various embodiments of the inkjet printing system 500, the substrate support device 520 may be a floating table provided for frictionless support of the substrate 522. In other embodiments, the substrate support device 520 may be a chuck. The Y-axis motion system 524 shown in FIG. 5, such as an integrated rail system, can use a linear air bearing motion system or a linear mechanical bearing motion system, and the air bearing motion system is a substrate 520 in the Y-axis direction. ) To enable frictionless transfer. In addition, the Y-axis motion system 524 can optionally use double rail motion, that is, provided by a linear air bearing motion system or a linear mechanical bearing motion system.

The camera assembly 560 as shown in FIG. 5 may include a camera 562, a camera mount assembly 564 and a lens assembly 568. The lens assembly can be mounted to the Z-axis movement plate 542A through the camera mount assembly 564. Camera 562 is a non-limiting example that can convert an optical image into an electrical signal, such as a charge-coupled device (CCD), a complementary metal-oxide-semiconductor (CMOS) device, or an N-type metal-oxide -Can be any image sensor device, such as a semiconductor (NMOS) device. As shown in FIG. 5, the Z-axis motion plate 532A can control the Z-axis position of the camera assembly 560 relative to the substrate 522. During various processes, such as, for example, printing and data collection, the substrate 522 can be controlledly positioned relative to the camera assembly 560 using the Y-axis motion system 524. The camera assembly 560 as shown in FIG. 5 can be controlledly positioned relative to the substrate 522 through X-axis movement through a carriage assembly 530B mounted on the bridge. Thus, the split-axis motion system of FIG. 5 accurately positions the camera assembly 560 and the substrate 522 relative to each other in three dimensions to capture image data on a portion of the substrate 522 at a desired point and/or height. Selection can be provided. As previously mentioned, other motion systems, such as the gantry motion system, can also be used to provide accurate motion in three dimensions relative to the substrate, such as between the printhead assembly and/or the camera assembly. In other embodiments, the motion system may position the camera assembly 560 relative to the substrate 522 using continuous or stepped motion or a combination thereof to capture one or more series of images of the surface of the substrate 522. have. Each image can include areas associated with one or more pixel wells, individual pixels, or combinations thereof, including peripheral surface areas (eg, including associated electronic circuit elements, paths, and connections). Any variety of cameras are used, which can have a relatively large field of view (>0.5 mm), providing very fine granularity (3 microns per pixel resolution).

FIG. 6 is a schematic diagram of system 600 as a gas enclosure that can accommodate a printing system, as previously mentioned. The gas enclosure system 600 can include a gas enclosure assembly 650, a gas purification loop 630 in fluid communication with the gas enclosure assembly 650, and at least one temperature regulation system 640. Additionally, various embodiments of a gas enclosure system may include a compressed inert gas recirculation system 660, which operates a variety of devices in OLED printing systems such as substrate float tables, air bearing linear slider assemblies and linear air bearing assemblies. It can provide an inert gas to let. Various embodiments of compressed inert gas recirculation system 660 can use a combination of the two as a source for various embodiments of compressors, blowers, and inert gas recirculation systems 660. Additionally, gas enclosure system 600 may have a filter and circulation system (not shown) inside gas enclosure system 600, which includes a cable tray exhaust assembly, a substrate floatation table, an air bearing linear slider assembly, and a linear air bearing assembly. Together with other configurations such as, it is possible to provide a substantially low-particulate printing environment.

As shown in FIG. 6, for various embodiments of the gas enclosure assembly 600, the gas purification loop 630 extends from the gas enclosure assembly 650 to the solvent removal configuration 632, after which the gas gas purification system ( 634) may be included. Then, the purified inert gas of the solvent and other reactive gas species such as oxygen and water vapor are returned to the gas enclosure assembly 650 through the injection line 633. In addition, the gas purification loop 630 may also include suitable conduits and connections and sensors such as oxygen, water vapor and solvent vapor sensors. A gas purifying unit such as a fan, blower or motor may be provided separately or integrated (eg, within the gas purification system 634) to circulate gas through the gas purification loop 630. Although solvent removal system 632 and gas purification system 634 are shown as separate units in FIG. 6, solvent removal system 632 and gas purification system 634 may be housed together as a single purification unit. Each temperature regulating system 640, for example, at least one cooler 641 (e.g., may have a fluid discharge line 634 to circulate refrigerant into the gas enclosure assembly) and fluid injection to return refrigerant to the cooler Line 645 may be included.

For various embodiments of gas enclosure assembly 600, the gas source can be an inert gas such as nitrogen, any inert gas, and combinations thereof. For various embodiments of the gas enclosure assembly 600, the gas source can be a source of gas, such as clean dry air (CDA). For various embodiments of gas enclosure assembly 600, the gas source may be a source that supplies a combination of inert gas and gas such as CDA.

In addition, system 600 helps maintain levels for various reactive gas species, as desired, wherein the reactive gas species are water vapor, oxygen and/or 100 ppm or less (eg, 10 ppm or less, 1.0 ppm or less) , Or 0.1 ppm or less). In addition, the gas enclosure system 600 helps provide a low particulate environment that meets a range of specifications for particulate matter carried in air, in accordance with ISO 14644 Class 1 to Class 5 clean room standards.

Note that, as previously mentioned, gas enclosure system 600 supports printing and/or imaging. For example, a gas enclosure system can use a substrate floatation table, an air bearing linear slider assembly and/or a linear air bearing assembly to move the substrate into and out of the printer. In addition, as discussed above, the gas enclosure system can optionally wrap a data collection device (eg, camera, camera pointing system and other image capture elements).

7A shows a flowchart 701 associated with one method for processing image data. This method can optionally be implemented as a command language stored on a computer-readable medium, ie as software or firmware controlling one or more processors, as previously introduced. As shown by number 703, the captured image data is first received by the processor. Such data could be captured under the control of the processor, or recovered from files stored locally or remotely. In one embodiment, captured image data is received directly from the camera when capturing each image as a “stream” of the captured image, while the camera is scanned continuously or cascaded over various pixel wells. The captured image data is then filtered (705) to separate the deposited film (or films) of interest. For the filtered data generated as a result of this process, for example, it represents image data associated with the interior of one or more pixel wells, and then film integrity is analyzed (707). It is noted that in one embodiment, this analysis was previously mentioned that the finished layer is performed after any post-printing completion step. In another embodiment, this analysis (and image capture) can be performed on still-wet ink to identify situations where filling is incomplete, air bubbles are present in the ink, or there is another potential defect. In various embodiments, a healing effect 708 can be taken in response to a detected defect. For example, the OLED device substrate can be discarded (and thus potentially saves additional time and deposition costs), or alternatively, the defect is by removing the deposited layer, by adding the deposited layer, or deposited The layer can be fixed by further treatment. If the analysis is performed on a wet ink, for example, the wet ink can be removed and the deposition re-performed. If air bubbles are present, the ink is heated or otherwise processed to remove the air bubbles. If the filling is incomplete, additional ink is applied to the correct position to complete the filling. There are many other possible remedies that provide a substantially more efficient manufacturing process.

If no defect is identified, processing continues to the next pixel well (709). These functions are shown in dotted lines to indicate optional properties. In embodiments in which layers within a plurality of wells are overhauled at one time, a given captured image may represent more than one pixel well and more than one deposited layer needed for analysis. The number 709 should be the case, and the captured image received in step 703 can be further processed for this additional pixel well (if any). In one embodiment, each captured image includes an area associated with one or more pixels, and in the second embodiment, each captured image contains only a single pixel well (or filtered for verification), and a third embodiment In, multiple pixel wells representing one or more pixels or a subset of one or more pixels may be processed at a time. If a given captured image has been completely processed, or if additional pixel wells still require analysis for a given deposition, the method loops to think of the different captured images denoted by numbers 711 and 713. When all pixel wells have been evaluated and passed the desired quality threshold, i.e., no defects are detected, the analysis is complete, as indicated by number 715.

FIG. 7B is used to provide additional detail for an exemplary filtering operation, which is used to separate image data corresponding to a layer deposited directly under scrutiny. This filtering operation is usually indicated by a series of steps 721. As indicated by numeral 723, image data corresponding to the captured image is received and stored in a fast-access memory such as RAM. Such an image can be received immediately or recovered from a machine-readable medium, as mentioned above. Note that FIG. 8A provides an example of one virtual image, showing two and a half pixel wells in one view. In one embodiment, processing is to focus on one of these pixel wells at a time, but in another embodiment, multiple pixel wells (eg, wells 807 and 809) can be considered simultaneously. A boundary detection function 725 is applied to separate the expected mitt from the bound of the deposited layer under scrutiny. Optionally, this detection function can be applied using a gradient function, that is, it allows to identify a boundary or contour corresponding to the periphery of the layer (727). For example, these contours can be compared to expected data (729) to determine the accuracy of the filling. For example, looking at FIGS. 8A-8F, the presence of the region 832 of incomplete filling is achieved by comparing the data for the reference pixel well (FIG. 8C) and the filtered image data for the increased layer (FIG. 8D). It can be detected to identify any type of deposition error or discrepancy. Depending on whether the layer metrics and bounds are compared to the reference data in this regard, the identification of pixels through image analysis (eg, gradient analysis) can be used to create a mask 731, the image corresponding to the content of the pixel well It can be used to separate data. This mask is represented by number 825 in FIG. 8D, and the mask is applied to filter the captured original image, producing filtered data that separates (733) the content of one or more pixel wells under analysis. Then, the filtering operation ends (735) for the given well in the given image.

FIG. 7C is used to discuss the use of a gradient function to identify discontinuities in image data representing the enhanced layer and processing to filter data, including optional boundary processing, introduced above with reference to numbers 727 and 729. do. This processing generally occurs in a series of steps, identified by number 741 in FIG. 7C. As with other flow charts discussed herein, the method shown in FIG. 7C may be performed again by one or more processors operating under the auspices of instructions stored on a computer-readable medium. With regard to the processing of FIG. 7A, filtered image data is possible per well, each to evaluate the quality of the deposited film in a given well.

For filtered data, image features are optionally used per process 743, to evaluate the quality of the separated and discussed layers. Such image features can include, without limitation, color intensity, hue, brightness, or other features. In one embodiment, although further discussed below, color data is converted to brightness values according to well-known formulas for summing the intensities of a plurality of color components, such as to obtain grayscale intensity. For a given color value only, other possibilities for focusing processing on color difference values or other measurement values are possible. The highlighted data representing the desired characteristics (or filtered data), in an optional process for detecting underfill, is in accordance with the numbers 745 and 747, in the expected pattern for the layer within a particular print well (eg, the expected contour). It is compared with the corresponding stored file data. For example, as mentioned with reference to FIGS. 8A-8F, the selected features can be compared to outline data from a stored file representing the expected contours of the deposited layer. As part of this comparison, the image processing software performs contour analysis to see if there are errors, if the errors are higher than the level of quality or filling defects (749), and whether the special remedy should be viewed as any identified defect (751). Can decide. Note that any type of error measurement or comparison process can be used. For example, if there is an error in the contour of the deposited layer, but the error is below a given reference or set of criteria or area thresholds, the pixel well contour may still be acceptable. Citing another example, some applications may not know about the deposited layer contour that exceeds the expected well boundary, for example, errors can only be related to a lacking area. Many other possible criteria can be used. If the contour (eg, boundary) of the deposited layer is acceptable, the deposited layer can be analyzed for gradients within the deposited layer (ie, for discontinuities that may indicate pixel performance problems).

As part of this analysis, a gradient function is applied to the image data (ie filtered or emphasized data), confirming non-uniformity in the deposited layer, for number 753. In one embodiment, the gradient filter essentially acts as a high-pass filter, such that the lack of spatial variation results in a zero value for each PEL, while the gradient (or discontinuity) is a non-zero attribute (e.g., adjacent). This results in a difference in intensity between the 8-bit intensity values for the PEL, for example 8-bit absolute values, 0-255). This requirement is not the case for all embodiments; for example, processing can be used to provide zero or negative value output to indicate discontinuities, but there are obviously many variations. The result of the gradient analysis is processed data containing a set of values, where each value indicates whether a gradient or other discontinuity is within the dimensions of the selected image characteristic. This is indicated by the number 755 in FIG. 7C. Almost any error measurement can be used once again to determine if there is a problem such as interlayer delamination or another practical problem (757). If such a problem exists, once again, it may result in further processing or otherwise leading to another form of healing (759). If all image data for a given pixel well has been processed, the method proceeds to the next pixel well 755.

7D is a flow chart that identifies, in one embodiment, additional specific processing related to gradient analysis and quality analysis related thereto. The method of FIG. 7D is generally designated 761. As indicated by the number 763, the filtered image data is optionally simplified, as mentioned above, to highlight the selected feature (e.g., intensity as mentioned above), i.e. any desired image attribute. This can be used, or you can simply work on the raw or filtered image data itself. The selected data for processing (G) is extracted as a series of overlapping tiles and convolved with a gradient operator (represented by matrix specifier F) for process 765 and function 767, which results in processed data. In one embodiment, very simple operators are used to track very high frequency variations and essentially act as a high-pass filter between adjacent PELs. In this regard, the operator can be a matrix that convolves each PEL with its immediate neighbors in the filtered data (or the highlighted features of that data, such as intensity), and produces values that depend on the differences. do. In one embodiment, the resulting value is zero for continuous image data, and a non-zero value if there is a difference (discontinuity) from one or more immediately adjacent PELs. The result of such an operation is typically a set of processed data 769 (eg, a gradient value for each PEL represented by the result value or filtered data, where each gradient value is a discontinuity or a gradient). Emphasizing existence and quantifying the magnitude of this discontinuity). Then, these various gradient values are processed to detect possible quality problems associated with layers in the pixel well under consideration (771). Various measurements can be used to make this assessment. For example, as seen on the right side of FIG. 7D, and as indicated by the number 773, in one embodiment, the absolute values of the gradient values can be summed together, and the summation is compared to a threshold. For example, for all wells under consideration, the unsigned gradient value can be compared to an experimentally-determined reference value (eg, "250000"). If the total gradient size is less than this reference value, the pixel well can be determined to be healthy (eg, pass), otherwise the pixel well can be determined to have a problem.

Obviously, many other criteria may apply. As a second option represented by box 775 in FIG. 7D, the gradient value can be compared to one or more statistical measurements, for example, as indicated, the variance or standard deviation of the gradient value can be compared to a threshold (eg, If the variance or standard deviation is small compared to the threshold, the pixel well is "passed" for a particular layer deposition). Obviously, many examples are possible, including any statistical test form 779, as represented by the box labeled “Other Criteria” 777 in FIG. 7D. Note that a combination of these measurements can also be used. In one specially-considered embodiment, the sum of the gradient size and statistical measurements (e.g., variance or standard deviation) for a given pixel well is calculated and compared to a threshold, so that both sum and variance are less than each threshold. If it is, the deposited layer is "passed".

At number 781, if it is determined that the layers have all passed for the pixel wells associated with the deposition of a particular layer, processing of the OLED device substrate continues with the EH other pixel well layer, panel encapsulation layer or other process. Alternatively, if a problem is detected, processing stops, and a remedy is attempted to fix the identified problem and save the OLED device substrate.

8A-8F are used to further illustrate the various processes discussed above.

8A shows one fictitious example of captured image 801. It can be seen that such an image includes three pixel wells formed on the substrate 803, including portions of the first pixel well 805, the second pixel well 807 and the third pixel well 809. . These various pixel wells can form part of the circuit for a single pixel in an OLED device, but this is not necessary. As indicated by numeral 811, each pixel well can be associated with an electronic device and traces used to control light generation by a light emitting layer within a corresponding one of the pixel wells. Of the three pixel wells, the middle well 807 has a hypothetical error filling 832 (as a deposited layer that has not reached the bank in this region), and as referenced by the number 835, the deposited layer It has defects within the area.

In contrast, FIG. 8B shows an optional step of trimming the image data to focus on one pixel well at a time, ie, the middle well 807 of FIG. 8A. In one embodiment, multiple pixel wells can be analyzed simultaneously (eg, if these wells receive the same layer material through a piping process), and in another embodiment, only one well may be processed at a time. Note that there is. In FIG. 8B, the image data 815 includes a well region 807, a bank structure 816 serving to trap fluid deposited in the well, and, for example, structures and electronics 811 outside the confines of the bank structure. The excess image data 813 shown is shown.

As previously mentioned, in one embodiment, the captured image data as represented by FIGS. 8A or 8B is filtered to remove excess image data. In one embodiment, this filtering is performed using boundary analysis, such as a second gradient function or other processing to detect the bank structure 816. In another embodiment, the boundary of the deposited layer is identified and compared to the reference data (indicated by reference number 822 in FIG. 8C) and used to determine if there is a filling defect. Regardless of the specific method, these operations can be used to create a mask to be used for this purpose (the mask is shown in FIG. 8D by numeral 825). Image data matching the hatched area 827 (eg, in FIG. 8B) is masked out and not used, and image data matching the clean area 829 is not filtered and will be used in subsequent gradient analysis.

The result of the filtering operation is shown in Fig. 8E by reference numeral 831. Generally speaking, in this embodiment, the filtered data is image data corresponding to the pixel well confinement (ellipse 807), null data in other areas (eg, number 833), and numbers 832 and 835. It contains variable image data corresponding to the potential defects indicated by. In this regard, an optional analysis can be performed to identify filling problems as indicated by the number 832. For example, if the method is applied to subsequent deposition of a wet ink (eg, opposite of a finished, cured layer), the area 832 can have substantially different color characteristics that can be identified using a color filter. And this applies in one embodiment. Depending on the nature of the defect, a remedy can be applied to heal the problem. It is noted that the number 835, which indicates a possible defect in the deposited layer, may have the general color characteristics of the deposited layer, so it may remain unidentified at this stage, depending on the nature of the processing applied.

8F shows processed data 841, which is generally the result of application of a gradient filter (not shown). In these figures, the value corresponding to each PEL (e.g., FIGS. 8B-8E) corresponds to the discontinuity 842 or 845, but these values appear to have empty values, such as gradient outlines or nested sets of outlines. Is expressed as The dash line 843 is superimposed on this figure, so as to simply check the pixel well contour. Prior to the application of the gradient filter, the layer discontinuity 835 may remain undetected, while the result of applying the gradient filter intensifies differences in strength that may indicate problems such as interlayer peeling of one or more OLED stack layers.

9-10 show histograms 901 and 1001 of the absolute value of the gradient represented by the processed data for two virtual pixel wells, respectively. Descriptions corresponding to pixel wells can be seen on the right side of each figure, numbered 903 and 1003, respectively. In each figure, the numbers 904 and 1004 indicate the location of the pixel confinement bank. Each description 903 and 1003 represents a gradient value corresponding to PELS representing an image of the associated pixel well, so 903 and 1003 represent discontinuities in the captured image data associated with the depicted pixel well. 9 shows the data for the deposited layer 905 without quality problems, while FIG. 10 is actually expressed by quality problems, such as interlayer delamination or the “racetrack effect” that occurs in the center of the Fissel wells. It shows data for another defective deposited layer 1005 in the center of the pixel as shown. While the gradient described in this example can be taken as a grayscale intensity value, in practice, the gradient can be applied to many other types of data representing or derived from the captured image.

As indicated by the figure, referring to the histogram 901 of FIG. 9, the “y-axis” of the histogram represents the number of occurrences of a given gradient size, while the “x-axis” represents the gradient size. As can be seen in the upper right corner of the figure, in this example, the sum of the absolute values of the gradient is "164,391", and the pixel well description 903 can be seen to correspond to the layer deposition without any visible quality problems. . Note that, as indicated by the histogram, there is no significant number of occurrences of large gradients, and the histogram has a 3σ gradient value of approximately “21”. Approximate values for the quality layer in the pixel well are: layer type (and resulting image appearance), camera type, magnification, lighting, distance from the substrate, whether the layer is wet or dry/cured, pixel well size and many other It can vary depending on the element. One or more thresholds for accessing layer quality are typically determined experimentally. In this example, the summation will be compared to a threshold (eg, “250,000”) determined with respect to a particular manufacturing process, OLED device and machine (including camera). In one embodiment, intentionally-infused pixel defects using specific equipment, such as using a prepared substrate with good and poor pixel fill (ie, layer defects), or accessing these thresholds Standards are prepared and analyzed in advance by dynamically printing (i.e., into a reference well) these standards. It should also be noted that, as mentioned previously, in one embodiment, multiple quality tests are applied. In other words, the gradient sum shown in the upper right corner of FIG. 9 may theoretically result from the predominance of a relatively small number of very large gradients or a large number of small gradients (the latter is actually in the histogram 901 of FIG. 9). Expressed by). Thus, in one embodiment, statistical measurements, in this case 3σ analysis, are applied to distinguish these cases. As indicated, the 3σ level in the case of FIG. 9 falls to approximately “21”, indicating that under analysis, approximately 99% of the layer material in the pixel well produced a gradient value of less than “21”, a relatively small value. Since the standard deviation (σ) depends on the number of PELs and the square of the gradient values represented by the histogram, if the gradient sums (164, 391) were generated by more occurrences of large-sized gradients, these values would be larger. Will be expected. Thus, in this case, the 3σ level represented by the histogram is compared to another threshold (eg, “40”), and since the pixel data for this pixel well is all satisfactory for the test, the analyzed layer within the pixel well is It is determined that it has passed the threshold quality. It is noted again that these various tasks are typically performed by software (instructions on computer-readable media) that controls one or more processors to perform these various tasks (including desired healing actions). Also, as mentioned previously, many other types of criteria and tests apply without limitation, based on any type of statistical test, standard deviation, 3σ level, variance, summation, sum of squares, absolute value or others. do. In one embodiment, multiple tests, such as classifying the quality of a deposited layer, may be applied, such as by matching one of a gradient value for a layer and a range of values (identified by one or more thresholds). .

10 shows data for a layer determined to have quality problems. First of all, note the "racetrack effect", which can be seen in the gradient values and indicates possible interlayer delamination (eg, of the cured or finished film leading to a post-printing processing step to complete the film). Considering the racetrack effect, the histogram 1003 shown represents a much more frequent occurrence of a large gradient value (eg, representing a gradient size of 30-70) corresponding to the “racetrack” shape shown in figure 1003. It should not be surprising. That is, as indicated on the right side of Fig. 10, the sum of the gradient (absolute) values in this case (339, 409) is significantly greater than the hypothetical threshold of "250,000". Applying the above test criteria for FIG. 9, a comparison of these values will indicate that the layer can have quality problems and, again, a remedy can be applied as an option. It is also noted that in this case the 3σ level represented by line 1007 immediately exceeds “50” and exceeds the virtual threshold of “40” used above with respect to FIG. 9.

11 shows another embodiment 1100 of a method for processing image data. Additional, fewer, or different operations may be performed depending on the embodiment. For example, the method 1100 of FIG. 11 may provide additional functionality beyond the ability to process one or more images. The order of presentation of the work of the method 1100 of FIG. 11 is not intended to be limiting. Thus, although some of the optional flows are presented sequentially, various tasks can be performed in various iterations, simultaneously, and in a different order than shown. Various embodiments of the method 1100 of FIG. 11, when executed by one or more processors, have computer-readable instructions that cause the processor to perform various functions shown in FIG. 11 to analyze image data representing pixel wells. It may be in the form of image processing or in the form of a printer control application.

For task 1102, the captured image data is received, for example, from a memory, camera, or another recipient. In a first example, a single image can be received at one time, and in another embodiment, multiple images provide a very-fast analysis of an OLED stack film in hundreds of pixel wells, including, for example, a flat panel display. It can be received by the ultra-parallel image processing used. Thus, the captured image data received at operation 1102 includes at least one pixel well comprising an OLED stack film. If the captured image data includes a plurality of complete pixel wells, each can be analyzed sequentially or in parallel from a common image. If each pixel well has a different layer material to be analyzed, such as a light emitting layer for light of a different color, analysis of the image data for these wells is appropriate for the layer under each process, thread, threshold or analysis. Other specifics can be used. In one embodiment, image data representing a stream of images can be processed in real time, image processing software tracks well locations, and in the captured image (in software) which wells have already been processed and which wells have not been (And the location of each of these wells). Image data can be captured as part of the printing process of the OLED device substrate or as part of a subsequent quality assurance process. In addition, the data can be displayed simultaneously on the display of a data collection device or imaging processing device or other printed for human-vision inspection.

As another option, image data can be stored in a database of a computing device and read from a database under the control of the automated image processing application of method 1100, so that it can be received by the image processing application for processing. . The user can optionally select image data for reading using, for example, a user interface presented under the control of the image processing application of method 1100. As will be understood by those skilled in the art, image data in which a portion of a flat panel display substrate is captured may be stored in various formats. In an illustrative embodiment, the image file can be stored as unprocessed RGB color pixel data.

In operation 1104, the filtered image data representing the OLED stack under analysis can be separated from the captured image containing excess data (ie, outside the pixel well). Several different algorithms can be used for this purpose (ie, extracting data corresponding to an OLED film or stack from a captured image). As previously implied, in one embodiment, a gradient filter (i.e., a second gradient operator) can be used to identify the periphery of the layer under analysis, so that only data corresponding to the deposited layer can be extracted. . Briefly, it is also possible to use a template corresponding to the pixel well terrain, and use that sample to mask the image under analysis. In another embodiment, task 1104 is combined with an image capture step (i.e., the image data is captured only on that layer, or filtered by the camera, reducing the amount of excess data before being received by the image processing system). .

In operation 1106, the filtered data representing, for example, a layer of divergent material (“EML”) of the OLED stack can be converted into a format that emphasizes certain image characteristics. In this case, the color image data representing the deposited layer is converted into grayscale intensity data according to the well-known formula / (Grayscale) = red *0.3 + green *0.59 + blue *0.11 . In the various formulas referenced below, the grayscale intensity values can generally be expressed as G _x,y , where x and y refer to the PEL position in the captured image. Note that this type of emphasis is purely selective, that is, it is possible to directly perform gradient analysis on color image data, and also to capture forged image data. Obviously, other alternatives exist.

In connection with operation 1108, the filtered data representing the pixel well (highlighted or not) is convolved with a gradient filter to produce processed data representing the gradient in the deposited layer. In one embodiment, the gradient filter is represented as a relatively simple matrix, and convolution is applied using a window in the PEL of the filtered image data. The mathematical expression is as follows.

G represents the gray scale intensity value of PEL

Θ is the convolution operator

[Fi F2] is a gradient filter mask; And

The right row is the resulting processed data.

The depicted convolution operator produces a gradient value that emphasizes the vertical line difference between the horizontally-adjacent PELs of the captured image data representing the layer under scrutiny. All perfectly uniform layers will have a zero filtered image matrix. In view of this, the summation of the absolute values of all the gradient values represented by the processed data can provide a criterion for evaluating the uniformity of the layers, and thus the OLED stack, under close scrutiny.

With regard to the selection of gradient filter masks, many different types of masks can be used. The gradient filter can be distinguished based on the applied mask. The exemplary mask shown above presents a dual symmetry along the horizontal axis. Other types of gradient filters can also be used. For example, in a Sobel mask, the horizontal axis divides the top and bottom by the same [1, 0, -1], while the vertical axis divides the mask by [1, 2, 1] and [-1 ,-2,-1 ]. Divide. Some exemplary gradient filter masks that can be used are shown in the table below.

Note that these gradient filters emphasize both vertical and horizontal discontinuities in the imaged layer.

In operation 1110, the method proceeds to analyze the quality of the deposited film under close scrutiny. In other words, the method can be fully automatic, for example, under the control of software. Summing up the absolute values of the brightness differences can generate a score for the evaluation of the uniformity of the layers in a given pixel well, and thus the quality of the deposited OLED stack in a given pixel well. In this example, the very low gradient summation is uniform and correlates with pixels without interlayer delamination.

However, as previously implied, the gradient sum score alone may not be able to determine direction for film quality issues. For example, elsewhere is uniform, a film layer having a large discontinuity in the center of the film may have the same gradient sum as a layer having a uniformly small gradient dispersed throughout the film.

Therefore, in one embodiment, statistical measurements, such as variance, standard deviation, histogram or other measurements of matrix elements, are computed to be fairly uniform with a film layer (which may represent a defective film) with images containing significant intensity changes. It is used to effectively differentiate from films (which may represent a satisfactory film) that contain small changes in strength.

The quality of the OLED stack film can be determined by evaluating both gradients or processed data (eg, summation of processed gradient values) and data representing the distribution of gradient values across pixel wells. In various embodiments, data generated from the filtered intensity data can be compared to a threshold. Factors that can affect the quality of the OLED stack film image, such as the ink used to form the various lights, different positions of the camera, and different substrate structures and pixels, are again what constitutes good gradient summation and good histogram characteristics. It can affect once. Thus, using the standard, calibration can be used to define good gradient summation and statistical characteristics under a specific set of conditions, and threshold gradient summation and threshold statistical criteria can be determined by using an image of a layer of film with defined uniformity. have. Using simple statistical quality control, the mean and standard deviation for an acceptable pixel image can be determined. Depending on the level of quality desired, the error threshold (or set of thresholds) can establish a corresponding number of standard deviations above the average identified for good pixel values.

In operation 1112, it is determined in what way the layer within the pixel well is defective at the identified pixel cell location. For example, a decision can be made if the threshold defined for the layer where sum and variance is satisfied is not satisfied. If the layer is said to be defective in some way, processing continues to operation 1114. If it is determined that the layer is not defective at the identified pixel well location, processing continues to operation 1116.

In operation 1114, a layer in the pixel well may be identified, at least potentially defective. For example, certain pixel cells and/or wells can be flagged for inspection that is visible by the inspector, confirming the decision. Notification to the inspector or operator is provided to the inspector using a speaker, printer, display, or by preparing or providing a text, email or voicemail, or by storing information related to the decision in the database for later inspection Can be. In addition, other forms of healing as discussed previously are possible. In one embodiment, the result of the presence of a quality problem affects subsequent processing and/or subsequent OLED device fabrication processes, such as stopping further image considerations. In another embodiment, through operation 1116, processing continues, so that an inspector or operator is notified of the presence of any potential defects, and these potential defects are later resolved when appropriate.

At task 1116, a determination may be made as to whether there is more image data to process. If it is determined that there is more image data to process, processing can continue to operation 1100. If it is determined that there is no more image data to process, the method 1100 can end. Statistics are calculated with respect to the number of defective pixels identified for the entirety of the OLED device to determine whether processing should continue, or whether the OLED device should be rejected, or other healing action should be taken.

Alternative embodiments of the system and method are shown in the flow chart shown in FIG. 12. In method 1200, the step of separating (filtering) the image data for the film layer to focus only the layer information within the pixel wells, such as highlighting data that captures captured image data, such as grayscale intensity and applies a gradient function It can be performed after the step of switching to. As before, the captured image data is first received with respect to step 1202. This data is then converted to grayscale intensity data with respect to number 1204 to highlight this selected image feature. For step 1206, a gradient filter is applied to highlight the borders and edges of the deposited layer, OLED confinement bank and other features seen within the captured image data. Regarding steps 1208 and 1210, once processed (gradient) data is generated, inverted data may be selectively generated to help separate the contours inside the pixel well. As an example, a target image or mask is created for each pixel well, serving as a mask to mask gradient values corresponding to excess PEL (ie, representing structures other than the film layer to be analyzed). As an example, after gradient analysis, the pixel well confinement bank can be represented in white (gradient, non-zero values), while other structures containing pixel wells can be represented in black. Then, this data is inverted and processed to assist in forming a mask, where the value outside the pixel well is represented by black (zero value), and the inside of the pixel well is white (binary value "1"). ). Then, with respect to step 1212, this mask can be applied back to the gradient value to separate the gradient values for the layer of interest.

, 여기서, l은 반전된 그래디언트 행렬이고, T는 타겟 행렬(마스크)이다. 이 식은 본질적으로 마스크를 반전된 그래디언트 행렬(즉, 프로세스된 데이터) 위에 위치시키고, 마스크와 픽셀 우물 위치는 정확히 정렬되는 둘 사이의 상대 위치에 대한 최대 결과를 생성한다. 이는, 왜냐하면, 상기 연산은 부울(Boolean) "앤드(and)" 연산에 적용되고, 마스크는 하얀("1") 또는 검정("0")인 입력을 가지기 때문이며, "앤드" 연산은 둘이 정확히 정렬될 때 최대치를 생성할 것이기 때문이다. 마스크의 적절한 위치를 알고, 그리고 나서, 픽셀 우물의 영역에 바로 해당하는 그래디언트값을 분리시키기 위하여 프로세스된 (그래디언트) 데이터로 마스크하는데 사용된다.More specifically, a target white image may be generated to form a pixel well region, a white value ("1") in the region corresponding to the pixel well, and a black value (" corresponding to other regions of the target image). 0"). Then, the inverted version of the processed data (gradient value) is converted to binary, combined with the target image according to the following equation:

, Where l is the inverted gradient matrix and T is the target matrix (mask). This equation essentially positions the mask over an inverted gradient matrix (ie, processed data), and the mask and pixel well positions produce maximum results for the relative positions between the two that are precisely aligned. This is because the operation is applied to a Boolean "and" operation, the mask has an input that is white ("1") or black ("0"), and the "and" operation is exactly what the two This is because it will generate the maximum when sorted. It is used to mask the processed (gradient) data to know the proper position of the mask, and then to separate the gradient values directly corresponding to the area of the pixel wells.

In the remaining steps 1216-1224, if the gradient values for the film layer are separated, the data is processed, as previously described, to determine the layer quality (for the presence of defects), as discussed previously. Likewise, it provides notification and/or healing action.

Note that there are various embodiments of the above. For example, a gradient function can apply any step of the process to the image data, whether the image data is filtered or converted. Also, although a relatively simple example of horizontal gradient computation is mentioned in relation to the above equation, more complex gradient functions can be applied, for example, without relying on matrix mathematics, computing vertical gradients, or gradients across PELs other than the nearest neighbor. Is to compute

In another embodiment, image analysis focuses on a deposition region that does not necessarily include a pixel well. For example, in a system that deposits a layer of film over a large area, imaging can be used on one or more discrete target areas that make up to evaluate the film quality area on a parceled basis. As described above, these embodiments are generally captured for each target region or parcel, e.g., under analysis, and a gradient function is applied to this image data (or derivatives thereof, filters, highlights or others) ), function to grow the processed data. Then, this data can be analyzed to find the presence of possible defects within a given target area. Other variations exist.

Methods described above may be implemented in computer-readable media (eg, software or firmware) or manufacturing systems for machines, printers, OLED devices, or other forms such as network machines.

Although the principles of the invention have been described in connection with specific embodiments, it should be clearly understood that these techniques are illustrative only and are not intended to limit the scope of the invention. What has been described herein is provided for illustrative purposes and illustration. It is not intended to be exhaustive or to be exhaustive in describing the precise form described. Many modifications and variations will be apparent to those skilled in the art. The disclosed ones have been selected and described in order to best describe the principles and practice of the disclosed embodiments in the described fields, so that those skilled in the art can understand various embodiments and various modifications suitable for specific use. The scope of what is disclosed is defined by the following claims and their equivalents.

Claims (34)

When executed by a processor, the processor,
Receiving data representing a captured image of the deposited layer of film for a pixel well-the deposited layer forming a permanent layer of light-generating elements associated with the pixel well-,
By processing the received data,
Applying a gradient function to at least the data representing the layer deposited in the structural confinement of the pixel well, to generate a gradient value for each of the plurality of pixels in the pixel well,
Determining statistical measurements using gradient values generated for each of a plurality of pixels,
Compare with the determined statistical measurement and threshold,
Rely on comparison to detect the presence of interlayer delamination in the confinement of the pixel well,
A non-transitory computer-readable medium storing computer-readable instructions that, when the presence of interlayer peeling is detected, generate an output indicative of the presence of interlayer peeling in the structural confinement of the pixel well. The method of claim 1, wherein when executed, the instructions further cause the processor to process the received data to highlight image features associated with the deposited layer in the pixel well, and to apply gradient functions to the highlighted data to generate gradient values. , A non-transitory computer-readable medium having computer-readable instructions stored thereon. 3. The non-transitory computer-readable storage of claim 2, wherein when executed, the instructions further cause the processor to generate emphasized data by converting color-filtered image data to grayscale data. media. The non-transitory computer-readable medium of claim 1, wherein the gradient function comprises a filter selected from the group consisting of Sobel filter, Prewitt filter, and Kirsh filter. The non-transitory computer-readable medium of claim 1, wherein the gradient function comprises a high pass filter. The method of claim 1, wherein determining a statistical measurement comprises calculating an overall value as a sum of absolute values of gradient values of a plurality of pixels in a pixel well, wherein the presence of interlayer peeling when the total value exceeds a threshold is present. A non-transitory computer-readable medium having computer-readable instructions stored thereon. 7. The method of claim 6, wherein determining a statistical measurement comprises calculating a standard deviation value as a standard deviation of absolute values of the gradient values of a plurality of pixels in a pixel well, wherein interlayer separation when the standard deviation value exceeds a threshold. A non-transitory computer-readable medium having computer-readable instructions stored thereon, the presence of which is also detected. 7. The non-transitory computer-readable medium of claim 6, wherein each gradient value corresponds to an intensity difference between neighboring pixels of a plurality of pixels in a pixel well. The method of claim 1, wherein determining a statistical measurement comprises calculating a standard deviation value as a standard deviation of absolute values of gradient values of a plurality of pixels in a pixel well, wherein interlayer separation when the standard deviation value exceeds a threshold. A non-transitory computer-readable medium on which computer-readable instructions are stored, in which the presence of the is detected. The gradient function of claim 1, wherein the gradient function includes a matrix operator, and when executed, the instruction further performs a gradient function by the processor convolving the value of the matrix operator and the image data value representing the layer deposited within the confines of the pixel well. A non-transitory computer-readable medium having computer-readable instructions stored thereon. The method of claim 1, wherein determining a statistical measure comprises calculating a variance as a variance of the absolute values of the gradient values of the plurality of pixels in the pixel well, wherein the presence of interlayer delamination when the variance exceeds a threshold. A non-transitory computer-readable medium having computer-readable instructions stored thereon. 2. The non-stored computer-readable instruction according to claim 1, wherein the instruction, when executed, causes the instruction to further control the movement system configured to transport the inkjet print head and receive the captured image from the camera carried by the movement system. Transitory computer-readable medium. A camera for capturing image data comprising a layer deposited in a pixel well on a substrate, the deposited layer forming a permanent layer of light-generating elements associated with the pixel well-and,
At least one processor,
A computer-readable medium operatively coupled to a processor-a computer-readable medium having computer-readable instructions, which instructions, when executed by at least one processor, cause the apparatus to
Receiving the captured image data from the camera, and
By processing the received data,
Applying a gradient function to at least the data representing the layer deposited in the structural confinement of the pixel well, to generate a gradient value for each of the plurality of pixels in the pixel well,
Determine the statistical measurements using the gradient values generated for each of the plurality of pixels,
Compare with the determined statistical measurement and threshold,
Rely on comparison to detect the presence of interlayer delamination in the confinement of the pixel well,
And, when the presence of interlayer peeling is detected, to generate an output indicative of the presence of interlayer peeling in the structural confinement of the pixel well. 14. The method of claim 13, when executed, the instructions further cause the device to process the received data to highlight image characteristics associated with the deposited layer in the pixel well, and apply a gradient function to the highlighted data to generate a gradient value. , Device. 15. The apparatus of claim 14, wherein when executed, the instructions further cause the device to generate emphasized data by converting color-filtered image data to grayscale data. 14. The apparatus of claim 13, wherein the gradient function comprises a filter selected from the group consisting of Sobel filter, Prewitt filter, and Kirsh filter. 14. The apparatus of claim 13, wherein the gradient function comprises a high pass filter. 14. The method of claim 13, wherein determining a statistical measurement comprises calculating the total value as the sum of the absolute values of the gradient values of a plurality of pixels in a pixel well, wherein the presence of interlayer peeling when the total value exceeds a threshold is present. Device detected. 14. The apparatus of claim 13, wherein each gradient value corresponds to a difference in intensity between neighboring pixels of a plurality of pixels in a pixel well. 14. The gradient function of claim 13, wherein the gradient function includes a matrix operator, and when executed, the instruction further performs a gradient function by the device convolving the value of the matrix operator and the image data value representing the layer deposited within the confines of the pixel well. Device to be applied. 14. The method of claim 13, wherein determining a statistical measurement comprises calculating a standard deviation value as the standard deviation of the absolute values of the gradient values of a plurality of pixels in a pixel well, wherein interlayer separation when the standard deviation value exceeds a threshold. The presence of the device is detected. 15. The method of claim 13, when executed, the instruction conveys the inkjet print head and receives the captured image from the camera, the camera further controls the movement system configured to be carried by the movement system, wherein when executed the instruction And allow the device to capture at least one image of each pixel well carried by the substrate. A method of determining whether a layer of material deposited within a pixel well of a display panel meets a predetermined quality criterion, wherein the deposited layer forms a permanent layer of light-generating elements associated with the pixel well- silver,
Receiving captured image data of the deposited layer,
Processing the received data by a computing device −
Applying a gradient function to at least the data representing the layer deposited in the structural confinement of the pixel well, to generate a gradient value for each of the plurality of pixels in the pixel well,
Determine the statistical measurements using the gradient values generated for each of the plurality of pixels,
Compare with the determined statistical measurement and threshold,
Rely on comparison to detect the presence of interlayer delamination in the confinement of the pixel well-Wow,
And when the presence of interlayer peeling is detected, generating an output indicative of the presence of interlayer peeling in the structural confinement of the pixel well. 24. The method of claim 23, wherein the display panel forms an organic light emitting diode ("OLED") display panel. 24. The method of claim 23, wherein the layer forms a layer of an organic light emitting material. delete delete delete The method of claim 1, wherein the data representing the captured image of the deposited layer represents more than one pixel well, and the instructions further receive data for the processor and received for each pixel well represented by the captured image. Non-transitory computer-lead stored computer-readable instructions that cause data to be processed and instructions further cause the processor to generate an output indicative of the presence of interlayer delamination for each pixel well where the presence of delamination is detected. Medium available. 14. The method of claim 13, wherein the data representing the captured image of the deposited layer represents more than one pixel well, and the instructions further include that the device receives the data and is received for each pixel well represented by the captured image. Let the device process the data, and the instructions further cause the device to generate an output indicative of the presence of an interlayer release for each pixel well where the presence of an interlayer release is detected. 24. The method of claim 23, wherein the data representing the captured image of the deposited layer represents more than one pixel well, the method receives the data, and processes the received data for each pixel well represented by the captured image. And further comprising generating an output indicative of the presence of an interlayer release for each pixel well in which the presence of an interlayer release is detected. 24. The method of claim 23, wherein determining a statistical measurement comprises calculating the total value as the sum of the absolute values of the gradient values of the plurality of pixels in the pixel well, wherein the presence of interlayer peeling when the total value exceeds a threshold is present. Detected method. 24. The method of claim 23, wherein determining a statistical measurement comprises calculating a standard deviation value as a standard deviation of absolute values of the gradient values of a plurality of pixels in a pixel well, wherein interlayer separation when the standard deviation value exceeds a threshold. The presence of the method is detected. 14. The method of claim 13, wherein determining a statistical measure comprises calculating a variance as a variance of the absolute values of the gradient values of a plurality of pixels in a pixel well, wherein the presence of interlayer delamination when the variance exceeds a threshold. Device detected.

Images